by ERIC WEISS, PH.D.
In the October 1982 of Technical Analysis, I discussed Autoregressive-Integrated Moving Average
(ARIMA) and why it is such an accurate forecasting technique. In this article, I'll discuss how to build an
ARIMA model using typical software and present some examples from my program for microcomputers
such as the Apple ][, EASI/ ARIMA. Most other ARIMA programs running on large mainframe
computers use a similar technique, although the commands are a little different.
The approach I like best was developed and popularized by Box and Jenkins at The University of
Wisconsin and The University of Lancaster (England), respectively. They have had such a major role in
ARIMA that it is frequently called Box-Jenkins. Their approach identifies an appropriate model,
estimates it, runs diagnostic to make sure that it works as it should and then uses the model for